Development of ros applications for robust and safe operation of humanoid robot in dynamic environment

Cilj diplomskega dela je razvoj, testiranje in uporaba ROS aplikacij na humanoidnem robotu Pepper z uporabo že obstoječih knjižnic v okolju ROS. V testiranje so bile zajete funkcije zajema in obdelave slike iz kamer robota, ter uporabo obdelanih informacij iz kamer za varno vodenje robota v realnem okolju. Namen tega dela je čim bolj učinkovita in varna interakcija med robotom in realnim okoljem ter objekti v njem. Dokument opisuje funkcije in metode, ki se uporabljajo v ROS za pridobivanje informacij iz robotovih senzorjev ter kako in na kakšen način jih obdelujemo, da krmilimo gibanje robota. V delu je predstavljena aplikacija sledenja humanoidnega robota Pepper na podlagi obdelave slike v okolju ROS, ki je izvedeno na platformi Docker.The goal of the bachelor thesis is the development, testing, and use of ROS applications on the humanoid robot Pepper with the use of pre-existing libraries in the ROS environment. In the testing phase, we show capturing and processing functions of the robot camera, and the use of information from the camera for safe and robust controlling in real-time environment. Purpose of that work is to get the most efficient and safe communication between the robot and the real environment and object objects in it. The document describes functions and methods used in ROS for capturing the information from robots’ sensors, and how data is processed and used to control the robot’s movement. Finally, a tracking application of humanoid robots is presented, which is based on image processing in the ROS environment and Docker platform